- Break up guys into non-resonant lengths
- Use non-conductive guys
- Crank-up towers (keep them lowered in high winds and when not being used)
- Very heavy duty free-standing towers
This leaves us with using guyed tower, and the choice of conductive or non-conductive guys. Steel is often the least expensive and worry free option in comparison to kevlar/aramid composites such as Phillystran and fibreglass rod. I know hams who use non-conductive guys and they are happy with their choices. This choice frees them to focus on interactions between antennas and not with the guys.
I looked closely at the pros and cons of each approach and how others deal with them in their stations. Based on that I chose to go with steel guys. This article is how I am going about the challenge of designing the guys to minimally interact with antennas on the tower. Soon enough my swimming holes will be converted into foundations and tower raising will commence.
Apart from the terminations at tower end and anchor end each guy segment boundary requires 1 egg insulator and 2 guy grips. Many segments are required to make the guys non-resonant for most cases, which adds up to a lot of guying hardware.
If you choose your supplier with care the cost of a steel guy broken into non-resonant segments is cheaper than non-conductive guys of equal strength. After considering cost one must also realize that there is substantial work involved in cutting and splicing the segments. When done properly a steel guy built in this manner will be no less strong than a single run of steel guy wire.
Do not use clamps, crimps or other methods of terminating a steel guy wire. Guy grips, though they may look questionable to the untrained eye, are state of the art. On a guy wire broken into many non-resonant segments the risk of disaster with lesser quality terminations is high. Guys do break. I've seen it firsthand. Do not take chances with guying hardware.
From experience and the large body of work published by others I enter this experiment with a range of expectations.
- Yagis at the top of the tower will have negligible interactions with the guys except, possibly, on 10 meters with the top one or two guy segments.
- Gain will be affected little by guy interactions since all the selected guy segment lengths are non-resonant on the bands of interest. It would take a resonant or near resonant guy wire segment to have a large enough mutual impedance to affect gain.
- Similar to gain, SWR should only show small deviations due to guy interactions. That is, I expect SWR (impedance) to be a poor indicator of pattern distortion.
- F/B and F/S will be degraded when the yagi is side mounted; that is, mounted below the top set of guys, even when induced currents on the guy wires is relatively small in comparison to the currents on the parasitic elements. It is the minor lobes of a yagi that are most susceptible to interactions since a finely tuned distribution of phase and current is needed for fields to cancel.
- Interactions are greatest on guys that approach being parallel (or in a parallel plane) to any yagi element.
- The most problematic interactions will be on 10 meters, modest on 15 meters and negligible on 20 meters. I am not testing 40 meters at this time since interactions are unlikely with the guy segment lengths I chose.
Guy wire segment lengths
The amateur radio literature has quite a lot of material on guy wire resonance. Some make bald statements about what lengths to use or avoid while others dive deeply into specific cases. Below are a couple of example charts. On the left is the ARRL Antenna Book and on the left is from an NCJ article by N2IC.
Without meaning to be unfair to the authors these charts exemplify what I said above. The ARRL chart implies that there are good lengths and bad lengths. The N2IC chart makes it seem that any length longer than 12' (4 m) for the 10 meter example are problematic, at least to some degree. The authors in both cases do provide deeper discussion about the issues, and that is very good.
Yet in my years of discussing this matter with many hams of my acquaintance the discussion is usually lost on them. They only remember the simplicity of a single chart. That drives their decision process, unfortunately. I hope to do a little better with my modelling effort.
The lengths I will use in my first model are from the aforementioned sources. My initial choices are, from the top of the guy: 5' (1.5 m), which is 4' plus half the tower width; 6' (1.83 m) for the next 2 segments; 19' (5.8 m) for the next 3 segments; and 43' (13.1 m) for the final segment. For the upper guys another 43' segment may be required, plus a final segment of variable length to the anchor, but this is not modelled since it is far enough away from the yagis to be of far lesser concern.
Note: These lengths do not consider the 12, 17 and 30 meter bands. The primary purpose of my tower is for contests, which does not include those bands. In any case coming up with non-resonant lengths for all HF bands is nigh impossible. If that's important to you I strongly suggest you use non-conductive guys.The reason the shortest lengths are at the top is because they are adjacent to antennas mounted directly above them -- either at the tower top or side-mounted. It is good practice to place large wind loads near guy stations. Short lengths of under 10' (3 m) have essentially no interactions on any HF band. As the guys go out farther from the tower they are more distant from the antennas and consequently have lower mutual impedance. By distant I mean with respect to wavelength, not an absolute measurement.
Building a model
A complete computer model of guys and antennas is excessively complex and large. It isn't strictly necessary. What we do need is the minimum to test interactions with antennas mounted at various heights, especially when side mounted, and varying orientation to the guys when the antenna uses a rotator.
My interaction model contains 3 guys of identical construction. They are broken into the selected lengths, joined at the top and angled downward. Guys are joined at the top since the first segment is tied to the tower and together they form a short inverted vee. This important factor is not addressed in some studies.
The tower itself is omitted since it is orthogonal and symmetric with respect to yagis on the tower, and other antenna types that have symmetry with respect to the tower (e.g. inverted vee), and thus has negligible interaction. The same is true of control cables and coax running up the tower. The tower appears virtually in the model by lengthening the upper guy segments to account for the tower width.
The adjacent plots include my first test antenna. The intent is to have the guy wire model and then import and position an antenna model from my large library of EZNEC models. The antenna is easily moved up and down, either above or below the guys, the pattern and SWR generated and compared as secondary traces on one azimuth or elevation far field plot. The guys can be rotated to simulate antenna rotation (this is easier in the model than rotating the antennas!) to find best and worst case scenarios.
Antennas that are mounted below the guys simulate side mounted yagis. The baseline plot is taken with the antenna well above the guys to serve as the baseline for the comparison; that is, where the interactions are negligible.
Constraining the model
Although the tower will have 4 sets of guys only one set is in the model. This is adequate for antennas above the tower and above the next to highest guy station. Lower antennas will point through 2 sets of guys. Since I do not plan any important antennas lower than halfway up the tower this is the most that needs to be modelled.
An important consideration is the number of segments in the model which is pushing the limits of the software and my patience in waiting for each run to complete. The time needed to alter the segmentation of the guys to experiment is multiplied by doing all 12 guys rather than one set of 3. Since I can identify problem areas with just the one set I do not strictly need to model more. I can simulate that, when desirable, by lowering the antenna so that the guy set appears to be a higher set.
My initial runs fix the angle of the guys with the tower at 45°. In the actual tower the angle for the guy sets is 40°, 49°, 60° and 75° starting from the top. In fact, this is approximately true for any tower with 4 equally spaced guy stations that follows the 80% rule -- anchors located 80% of tower height from the tower base. Therefore 45° is a good proxy for the upper two guy sets. Interactions for antennas side mounted on the bottom half of the tower will see increased interactions with the lower guys since they are more horizontal. This matters but is not my immediate concern.
A peculiarity of my model is that I chose to eliminate ground and model in free space. Ground does matter, though most often only to a small degree with regard to guy interactions. As with eliminating multiple sets of guys I do this to focus on the direct contribution of the guys to antenna behaviour. Too many variables make for a big muddle from which reliable conclusions can be difficult to obtain. If you want the model to completely predict behaviour you'll need to include ground and all guys. However, that model may obscure the most important problem areas you need to address.
Because I am modelling in free space I will only show azimuth plots, and those will be at 0° elevation. This is not what we will encounter in the real world. What it does do is identify guy positions, guy segment lengths and antenna orientations that are going to cause problems. That is what I want to discover.
For the initial modelling runs I used a small 3-element tri-band trap yagi that I developed a few years ago. This allows convenient interaction testing on the most susceptible bands with one antenna. Once problem areas are identified I will substitute antennas that are similar to the ones I intend to put on the tower.
There are in effect two variables to play with in the constrained model: antenna height and guy orientation. Here is how I approached both.
I selected four heights in this experiment, with respect to the 25 meters height, with the guy apex 50 cm below that (24.5 m): 13 meters above; 0 meters; 7 meters below; and 10 meters below. As mentioned earlier the first height is the baseline for comparison, a height that ensures negligible interactions. The second represents the yagi mounted directed above the guys, whether top or side mounted. The third case is a side mounted yagi that is tightly tucked underneath the guys. The final case is about as low a side mounted yagi could go without running up against the next lower set of guys.
I then chose three guy orientations that are representative of scenarios that I expected to be most illustrative of the spectrum of interactions: one guy directly in the yagi's forward direction (top of diagram); one guy directly to the yagi's rear direction; and one guy in a plane parallel to the yagi elements. My results seem to indicate that I chose these scenarios well.
What do you think? Before I disclose the results and my interpretations it promotes understanding to look at an unfamiliar problem to consider how it might be solved. I recommend spending a few minutes right now doing that. You're on the honour system so I'll assume you've done that and I can continue.
The second worst case scenario is the one on the far right. The rightmost guy is in a plane parallel to that of the yagi elements and so is more similarly polarized than the other guys. Not only that, the affects are asymmetrical. Pattern distortion is evident on both 10 and 15 meters.
The primary plot is for 15 meters height which is ~10 meters below the guy apex (as explained earlier). Again, all scenarios were run in free space to minimize variables; the heights are there for future use only when I may run the models over real ground.
Notice the degradation of F/B and F/S in these azimuth plots -- there are also distortions in the elevation patterns, which are not shown here. For the yagi directly above the guys (25 m) the pattern distortion is very slight. For this antenna and for this guy configuration there is some peril in side mounting, but it isn't dreadful.
Gain is almost unaffected (no more than 0.2 db reduction) and F/B and F/S though worse are still respectable. SWR is barely affected. When side mounting for stacking gain it should work out just fine.
The worst case is for the guy orientation in the centre of the earlier set of diagrams. Does this surprise you? It surprised me. I expected the one discussed immediately above (guy parallel to elements) to be the worst. Looking at it more closely I can see how I was perhaps misled in my expectations. I have not evaluated it in detail so I can only suggest why this is occurring.
Notice how much the F/B and F/S have deteriorated especially on 15 meters, and there is a significant gain drop at the lowest height. On 10 meters it is not so bad, and is arguably no worse than for the case with the parallel guy.
Let's first look at the leftmost orientation. You might think that for the sensitivity of interactions that determine F/B and F/S that guys behind the yagi would be problematic. Yet this isn't the case. I suspect the reason is much the same for why it is possible to tune a yagi by pointing it upward, with the reflector close to the ground. Field cancellation to the rear and sides that is typical in a yagi reduces the potential for interaction with guys in those directions. The EZNEC Current table supports this interpretation.
In the forward direction the presence of guys will interact since that is where the field is strongest. Currents in those guys will upset the balance of phase and amplitude responsible for good F/B and F/S (directivity). Again, the currents table supports this interpretation.
I not only plotted the patterns I also looked closely at the currents induced on all guy segments. In EZNEC use the Currents table; do not rely on the graphical view since small but significant currents are not visible. Where the current is negligible the interactions are of no consequence. In all cases where the patterns were distorted there are guy segments with currents of only 5% to 10% that in the yagi's director and reflector. That is enough to disturb the fine balance required for best directivity in an optimized yagi with 3 or more elements.
The short guy segments at the top were in every case not responsible for interactions. Not even the top segment that joins with its siblings via the tower fasteners. Only when the segment length grew to 19' (6 m) did significant currents appear in select orientations, heights and bands. In the case of one parallel guy it was the two 19' segments in that guy that exhibited significant current. In the other cases the 43' (13 m) segment lower down also developed significant current in the guys positioned ahead of the yagi. This appears to confirm the reasoning I described above though it is not certain.
Side mounting can be a problem with selected orientations and bands for guys broken into non-resonant segement. Non-resonance is no assurance of good pattern since even small currents can wreak havoc with directivity. All you can hope to do is manage the problem by careful engineering.
Yagis above the tower are largely unaffected by guy interactions. The closest guy segments are short and are far from horizontal orientation. This is even true on 10 meters.
For fixed side mounted yagis the guy (or guys) ahead of the yagis can be chopped into smaller segments to preserve the pattern. However the cost versus benefit is doubtful, especially since you might in future decide to add a rotator to that yagi.
A guy aligned with the yagi boom does not interact at all. This should not be a surprise since it is orthogonal to the elements. I expected this.
That 15 meters performance came out worst is not necessarily indicative of what will occur with long boom mono-band yagis or other antenna types. Longer booms put the outer parasitic elements closer to the guys, leading to greater coupling and guy currents. Yagis further down the tower can also be more affected by interactions since the guys below and around them are more horizontal that those higher up.
You may have noticed that I did not show the plots for 20 meters. The reason is that the pattern distortions due to interactions were small, even for the side mount cases. Gain deteriorated by no more than 0.25 db and F/B by no more than about 3 db. While these figures make for uninteresting plots the message is that my selection from the literature of guy segment lengths works well for 20 meters. At least that is the conclusion so far, in advance of more detailed modelling.
The model I've presented here and the ways in which I've exercised it is a good start though far from the complete story. I intend to perform additional modelling to evaluate at least the following:
- Lower guys that are more horizontal and that will therefore have a higher mutual impedance.
- Effects of looking through 2 sets of upper guys from a lower side-mounted yagi.
- Long boom yagis, which will get closer to the guys when side mounted.
- Fixed wire yagis where inverted vee elements will be more parallel to the guys.
- WARC bands, which while not a great concern I'd at least like to know what to expect when an antenna for 30, 17 or 12 meters is side mounted.
- Adjust guy wire segment lengths hopefully to tame interactions on 10 and 15 meters.
I can, and have pumped the water from my swimming holes to inspect them. Repairs to the damage winter has caused will have to wait for heavy equipment to arrive. They're a mess and every rainfall refills the holes. Indeed, we are experiencing extensive flooding in this part of the country. I had a near disaster yesterday when the sump pump failed overnight and the basement began flooding. Luckily I had an emergency pump handy -- the same I used to pump the tower excavations. The loss is minor but it sure is wet down there.
Whether it's guy wire interactions or sump pumps there is lots to keep one busy building and maintaining a large station located in a sparsely populated region.